STR: 1. Load http://www.spacegoo.com/helico/ (If it's still "Loading..." after ~10 seconds, reload. It should eventually give you a "START!" button.) 2. Click "START" button, when it becomes available. EXPECTED RESULTS: Terrain should be visible for your helicopter to navigate through. ACTUAL RESULTS: You're just navigating through a dark rectangular tunnel, with video game banners on the sides -- no terrain is visible. Eventually you'll crash into an invisible bit of terrain. Firefox 17.0.1 release gives EXPECTED RESULTS: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:17.0) Gecko/20100101 Firefox/17.0 So does Chrome Dev channel (Version 25.0.1364.5 dev) Nightly gives ACTUAL RESULTS: Mozilla/5.0 (X11; Linux x86_64; rv:20.0) Gecko/20121227 Firefox/20.0

Lots of errors here... on first load, there's NS_ERROR_FAILURE raised on what looks like a texImage2D -- they're probably not loading images correctly (not waiting until they're loaded). On reload, and during play, lots of: [14:28:00.475] Error: WebGL: DrawElements: bound vertex attribute buffers do not have sufficient size for given indices from the bound element array @ http://www.spacegoo.com/helico/:195 I'm guessing this is the reason for the missing terrain.

Treating this as a regression for now (since that's how it initially manifests, to users/webdevs), though per comment 1 it might end up being a Tech Evang bug. Regression range: { Last good nightly: 2012-09-21 First bad nightly: 2012-09-22 Pushlog: http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=48c4938eaf57&tochange=9cfb80a82883 } That pushlog contains several WebGL changes from bjacob --> CC'ing him. So -- this is a regression (modulo comment 1) in Firefox 18, which is being released soon...

(In reply to Daniel Holbert [:dholbert] from comment #4) > Pushlog: > http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=48c4938eaf57&tochange=9cfb80a82883 > That pushlog contains several WebGL changes from bjacob --> CC'ing him. For the record, those changes are: { > 94111544f201 Benoit Jacob — Bug 790946 - Remove support for the MOZ_ prefixed EXT_texture_filter_anisotropic ext name - r=jgilbert > ff0e65d0ee17 Benoit Jacob — Bug 732660 - Efficient drawElements validation on subarray and/or dynamically updated array - r=jgilbert > b8ee83073a77 Benoit Jacob — Bug 777967 - move the InvalidateCachedMinInUseAttribArrayLength calls a bit - r=jgilbert > cb2a1ae48b7a Benoit Jacob — Bug 790946 - Remove support for the MOZ_ prefixed EXT_texture_filter_anisotropic ext name - r=jgilbert > b8981694b0cc Benoit Jacob — Bug 790138 - raise WebGL context limits on desktop - r=jgilbert > b6ac16c3a5c7 Benoit Jacob — Bug 790086 - Try removing the Optimus WebGL/ANGLE blacklist - r=jgilbert What could possibly go wrong? > 5c5001289c36 Benoit Jacob — Bug 786383 - add MOZ_AVOID_OPENGL_ALTOGETHER option - r=karlt }

...and given the drawElements error message in comment 1, Bug 732660 (https://hg.mozilla.org/mozilla-central/rev/ff0e65d0ee17 ) looks like the most likely candidate for being the change that caused this.

I can't reproduce this locally. Is this bug still relevant?

Can't reproduce in latest Linux Nightly. (Tried both with + without e10s, for good measure.) --> WORKSFORME, hooray! Thanks, ashughes.